The predictive maintenance 'crystal ball'

Predictive maintenance programs provide miners with valuable insight into the condition of their in-service equipment, and are crucial in keeping mining and mineral processing applications online, streamlining maintenance schedules and minimising repair expenses.

Today's mining and mineral processing environment can be unforgiving, driven by the quest for quality, throughput and arguably most importantly, profit. With many mines operating 24 hours a day, seven days a week, it is crucial that their processes are kept online, and free from interruption. To achieve this, individual process equipment elements and machinery units must perform at an optimum level across a range of demanding operating conditions, all while remaining fault free. Here, the expectation from the mining sector has never been higher.

An unavoidable by-product of any mining and mineral processing application is the wear and tear on individual pieces of process equipment and their internal components. Breakages, leaks, overheating and complete breakdowns can lead to lengthy process shutdowns and lost revenue. These 'worst-case-scenario' breakdowns become more prevalent when process equipment is neglected or poorly maintained. To ensure process equipment and machinery are kept in optimum working order, a systematic maintenance program is vital.

The ability to accurately predict, and then address, the maintenance requirements of individual pieces of equipment goes a long way to maximising uptime and avoiding costly shutdowns. Here, predictive maintenance strategies lead the way. Offering 'crystal ball'-like insight into the condition of on-site equipment, predictive maintenance programs are fast becoming an essential component of modern mining processes.

React, plan or predict?

The ongoing maintenance of process equipment is a necessary operating overhead experienced across nearly every application in the mining sector. Equipment, such as motor gear-units, drive some of the mining industry's most vital processes and, as a result, need to be kept in good working order. This is often easier said than done.

According to SEW-Eurodrive applications engineer Luke Schmidt, gear-unit maintenance has traditionally been carried out three different ways — reactive, planned or predictive. "Reactive maintenance is the least efficient maintenance strategy," he says. "Addressing system faults and breakdowns after they occur means processes can be offline for long periods while spare parts are procured and repairs made. Usually, going offline for any period of time means loss of revenue, which is unacceptable. Furthermore, specialist service staff can be expensive, especially at short notice or unusual times."

Planned maintenance programs go some way to ensuring the wellbeing of gear-units, but can be viewed as unnecessary or wasteful, depending on the application. "Regular planned maintenance can be beneficial, but is often not very cost effective," says Schmidt. "It can result in unnecessary maintenance being carried out, which wastes time, materials, labour resources and money. Unwarranted oil changes and parts replacement, in particular, can have significant environmental impact." Further difficulties can be encountered, as planned maintenance activities are often carried out over summer shutdown periods — busy peak periods for maintenance staff.

According to SEW-Eurodrive Strategic Marketing and Product Manager Darren Klonowski, more sophisticated maintenance strategies, such as a predictive maintenance program, offer clear advantages. "Predictive maintenance allows businesses to monitor their process equipment, determine the condition and schedule maintenance accordingly," he says. "It ensures equipment life is maximised, and helps keep machinery online, preventing unplanned downtime."

Condition prediction

Predictive maintenance programs are centred around condition monitoring processes. Here, field-mounted sensors continuously detect and collect an array of performance parameters unique to individual machines. The collected data is then used to establish the 'real' condition of the machine and its components. This kind of insight allows maintenance technicians to take preventative action before a failure occurs, therefore avoiding the consequences of that failure.

"Condition monitoring permits the early detection of the initial stages of component wear which, if left unaddressed, can lead to catastrophic failures," says Klonowski. "This level of foresight is obviously a real benefit, particularly with respect to gear-units, which are often subjected to continuous operation under extremely demanding conditions. Being able to monitor the condition of bearings, gears, drive shafts and lubricating oil gives technicians a detailed understanding of the maintenance tasks that need to be carried out in order to extend the gear-unit life and keep it online longer."

Such field-mounted devices monitor a number of gear unit parameters, including vibration, oil level and temperature, and brake wear. "By monitoring these gear-unit parameters, on-site technicians can accurately plan maintenance activities," explains Schmidt. "It means maintenance is carried out only when and where it is required, which maximises component life and sees maintenance resources utilised efficiently."

Predict to protect profit

Predictive maintenance programs effectively address two primary considerations — they protect expensive process equipment, as well as entire critical process lines. "While there is merit in monitoring individual gear-units for the sole purpose of avoiding costly repair bills, often it is the protection of the process as a whole that is more important," says Klonowski. "The gear-unit being monitored may be relatively inexpensive, but it might be responsible for driving a process line with a high throughput rate. If this line goes offline, the mine can lose money at an alarming rate."

"It's not just the gear-unit that should be looked at," adds Schmidt. "It's the role of the piece of equipment in the process. The cost, production impact and flow-on effects associated with that piece of equipment going offline must be considered. A less-expensive gear unit might be responsible for driving a process that has a production output rate of $50,000 per hour. Predictive maintenance programs go a long way to ensuring these process stay online, while saving money."

According to Klonowski, predictive maintenance strategies are more cost effective when compared with reactive and planned maintenance schemes (see Figure 1): "Reactive maintenance results in periods of minimal maintenance followed by periods of extreme activity when urgent or emergency repairs are carried out — usually at an inflated rate," he says. "Planned maintenance delivers some cost benefits, but can result in unnecessary and costly maintenance being performed."

Figure 1: Predictive maintenance strategies are more cost effective when compared with reactive and planned maintenance schemes.

Detect, diagnose, act

At the heart of predictive maintenance installations lies a network of field-mounted condition monitoring devices or sensors linked via I/O to an onboard frequency inverter or central PLC. Each inverter or PLC can be connected to the site-wide SCADA system, and an internet mail server via an ethernet link (Figure 2).

"Such system architecture provides real flexibility," says Schmidt. "It allows machine performance and condition data to be easily detected, diagnosed and acted on. It can be configured to alert key personnel, once a predetermined alarm-point or milestone has been reached. The alarm can be sent to an on-site HMI, SCADA or PC, or alternatively to an off-site control centre, mobile phone or pager." Predictive maintenance strategies are especially valuable in remote locations where gear-unit breakdown and unplanned downtime can take extended periods of time to remedy. "By remotely monitoring applications in isolated or unsupervised areas, technicians can accurately predict maintenance requirements and plan ahead," says Schmidt. "It means maintenance resources can be deployed at a time that permits repairs and upkeep to be carried out on multiple pieces of equipment at the same location." In addition to being integrated into site-wide communications and control systems, many field-mounted sensors are equipped with onboard displays and indicators.

Figure 2: A typical condition monitoring network architecture.

"Visual indicators provide maintenance staff with an immediate indication of the gear-unit condition," says Schmidt. "Some devices actually show the 'hours to next service' of the gear-unit. More sophisticated devices, such as vibration analysis sensors, allow the early detection of roller bearing and gearing damage, as well as gear-unit unbalance and resonance problems."

As pressure mounts to keep mining and minerals processing applications online around the clock, predictive maintenance is emerging as a vital factor in efficient mining processes. With the ability to establish the current and future condition of in-service process equipment, predictive maintenance is set to become more than a maintenance option, but rather a necessity.

SEW-Eurodrive
www.sew-eurodrive.com.au
www.sew-eurodrive.co.nz